litellm/tests/e2e/load
Yassin Kortam 3d35eee560
test(e2e): derive the throughput SLO per replica and surface locust's error breakdown (#35494)
The floor was an absolute fleet number, so it asserted replicas x per-replica rate
and went red on how many gateway pods happened to be warm rather than on the request
path. The test now measures one replica first, with a short serial pass that only ever
occupies a single pod, and requires the concurrent phase to reach at least that rate.
A serial latency budget carries the request-path assertion the floor used to imply,
and both hold at one replica or seven.

Zero-error runs that "sustained 16.7 RPS" were queueing, not slow requests: the load
model is a mock_response deployment with no upstream, a single-worker replica serves
it in about 57ms, and 100 closed-loop users against 1/0.057 RPS of capacity sit at
6s each by Little's law.

The runner also kept locust's --json summary and threw away everything else, so a run
where 93% of requests failed said nothing about what they got. It now passes --csv,
reads the failure breakdown back, and reports locust's own generator-saturation
warnings, both folded into the assertion messages.

Resolves LIT-5054
2026-08-01 14:08:19 -07:00
..
conftest.py fix(e2e): register the load mock model through load_key instead of an autouse fixture so harness tests run without a proxy 2026-07-21 19:17:11 -07:00
load_client.py fix(e2e): migrate load suite from e2e_gateway to ProxyClient (#33839) 2026-07-18 19:41:03 +00:00
load_constants.py test(e2e): add Locust throughput load test that runs last (#33748) 2026-07-18 11:57:41 -07:00
locust_load.py test(e2e): derive the throughput SLO per replica and surface locust's error breakdown (#35494) 2026-08-01 14:08:19 -07:00
locustfile.py test(e2e): add Locust throughput load test that runs last (#33748) 2026-07-18 11:57:41 -07:00
session_anomaly.py fix(e2e): retry transient turn failures and move the weekly anomaly run to Saturday before the stable release cut 2026-07-21 16:55:44 -07:00
test_chat_completions_throughput_e2e.py test(e2e): derive the throughput SLO per replica and surface locust's error breakdown (#35494) 2026-08-01 14:08:19 -07:00
test_locust_load.py test(e2e): derive the throughput SLO per replica and surface locust's error breakdown (#35494) 2026-08-01 14:08:19 -07:00
test_session_anomaly.py test(e2e): move Admin UI Playwright suite to tests/e2e/ui (#34196) 2026-07-22 19:43:10 +00:00
test_weekly_session_anomaly_e2e.py fix(e2e): retry transient turn failures and move the weekly anomaly run to Saturday before the stable release cut 2026-07-21 16:55:44 -07:00
weekly_anomaly_config.yml test(e2e): add weekly session-anomaly load test against real providers 2026-07-21 14:54:02 -07:00